PROJECT SUMMARY This project relies on new biological insights into how SORL 1-retromer functions as an endosomal recycling unit to understand the structure and function of pathogenic variants in the SORL 1 gene that cause, or contribute to the risk of, Alzheimer's disease. Alzheimer's disease (AD) is the only major cause of death for which no preventive or significant disease-modifying treatment exists, and will cost the U.S. a trillion dollars annually by 2050. Only 4 truly causal genes have been identified: APP, PSEN1, PSEN2, and SORL 1. Of these, certain alleles of SORL 1 cause late onset AD that phenocopies the common sporadic disease. SORL 1 encodes a 2214-residue type I transmembrane protein that has 24 distinct domains. Its function in neurons is to help traffic membrane proteins such as the glutamate receptor subunit GLUA 1, APP, and other cargo to their correct destinations after endocytosis. To do this it forms a specific complex with cargo and with the retromer, a multiprotein assembly that is the master regulator of trafficking out of early endosomes. Pathogenic mutations cause endosomal traffic jams, ultimately leading to neurodegeneration. Since pathogenic SORL 1 mutations are involved in ~3% of all AD cases, it is very important to determine if a given mutation is pathogenic, and if the afflicted patient will be a candidate for therapy aimed at normalizing protein trafficking out of the endosome. Our overall approach is for human genetics to first identify likely pathogenic variants so that we can then more deeply probe their biology--- integrating structural, biophysical, biochemical, and cellular biology. This is innovative in expanding our mechanistic understanding into how SORL 1-retromer drives and modulates Alzheimer's disease. We will determine the effects of putative pathological SORL 1 variants on the structure, stability and cargo binding of SORL 1 protein. Both biochemical (dimerization, binding to the retromer multi protein complex, and binding of cargo) and biophysical (overall and local protein folding, protein stability, dynamical properties of the protein structure) will be measured. In the case of mutations in the VPS10/10CC portion of SORL 1, the crystal structures of selected mutant proteins will be determined to provide atomic-level details of the effect of the mutation on the local and global conformations of these two domains. In the case of mutations elsewhere in this large protein, we will use a model we have built (employing the deep learning algorithm AlphaFold2) of the ectodomain to aid in structure determination by cryoEM/ET, enabling us to predict the likely effects of mutations on the conformation - and biochemical functions - of SORL 1. This information will be incorporated, along with cell biology in neurons, plus histopathology, human genetics and mouse model data from other labs, to provide a comprehensive picture of the relationship between a given mutation's effects o...